How Many Companies Are Running Detectable UI Experiments, and Who Tests Most?
- By Ali Abouelatta
- n = 4,814
- Published
- Updated
- 2 min read
- experiments · mobile · web · saas · ux-patterns · monetization
Short answer
Lazyweb Research detected 4,814 before/after UI experiments across 362 distinct companies within its ~800-app corpus by July 2026; 2,160 of those experiments carry area-level annotations. Testing is concentrated: the most active company (Workday) accounts for 32 annotated experiments, and the top 20 testers each run 11 or more. The corpus skews mobile — 4,449 of 4,814 experiments (92.4%) are mobile. source[1]
Lazyweb Research detected 4,814 UI experiments across 362 companies; the most active tester ran 32 annotated experiments — July 2026.
Bar chart · Value
Value — The scale
One block is about 12.9 Value
- Detected experiments
- 4,814
- Companies with ≥1 experiment
- 362
- Annotated experiments
- 1,126
- Area annotations
- 2,160
- Mobile / web experiments
- 4,449 / 365 (92.4% / 7.6%)
The scale
Corpus headline numbers: source[1]
| Metric | Value |
|---|---|
| Detected experiments | 4,814 |
| Companies with ≥1 experiment | 362 |
| Annotated experiments | 1,126 |
| Area annotations | 2,160 |
| Mobile / web experiments | 4,449 / 365 (92.4% / 7.6%) |
The corpus is overwhelmingly mobile: web contributes only 365 experiments (7.6%), so most cross-cutting findings describe mobile behavior. source[2]
Who tests most
The most-annotated companies (distinct experiments): Workday 32, Navan 24, Memrise 22, YouTube Music 22, Slack 22, Microsoft Auth 20, Apple Music 19, Audible 18, ChatGPT 17, Plantin 15. source[3] The top 20 all run 11+ annotated experiments each. This is a long-tail distribution — a handful of very active companies plus a broad base of occasional testers across 362 firms. source[1]source[3]
How to apply it and caveats
Use this to calibrate benchmarks: when a page here cites a per-company or per-category pattern, the base is tens of experiments, not thousands. If you benchmark against 'what the field does,' remember it's really 'what these 362 tracked apps did,' skewed mobile. Caveats: these are observed UI diffs with inferred rationale, not measured A/B outcomes source[4]; company counts join a control screenshot to a company name, so labeling edge cases exist.
Real in-market screens
Screens from real, in-market apps, tracked by Lazyweb Research.
The numbers
| # | Stat | Computed from |
|---|---|---|
| [1] | 4,814 experiments / 2,160 annotations / 1,126 annotated experiments / 362 companies | statpack totals + companies_with_experiments |
| [2] | Mobile 4,449 (92.4%) / web 365 (7.6%) | statpack platform_split_experiments |
| [3] | Top testers: Workday 32, Navan 24, Memrise 22, YouTube Music 22, Slack 22, Microsoft Auth 20, Apple Music 19, Audible 18, ChatGPT 17, Plantin 15 | statpack top_testing_companies |
| [4] | Experiments are observed UI diffs with inferred rationale, not measured A/B outcomes | statpack methodology note |
Methodology
Universe: 4,814 detected before/after UI experiments (4,449 mobile / 365 web) across 362 companies in the ~800-app corpus, with 2,160 area annotations on 1,126 experiments, July 2026. Company names join a control screenshot; all counts are observations, not measured lift.
Sources
- [1]
Lazyweb Research analysis of 4,814 detected experiments across 362 companies, July 2026
2,160 annotations on 1,126 experiments; ~800-app corpus.
- [2]
Lazyweb Research analysis of 4,814 detected experiments, July 2026
Platform split: mobile 4,449 / web 365.
- [3]
Lazyweb Research analysis of most-annotated companies, July 2026
Distinct annotated experiments per company; top 20 each ≥11.
- [4]
Lazyweb Research methodology note, July 2026
Observed before/after diffs with inferred rationale, not measured results.
Source: Lazyweb Research — proprietary analysis of real, in-market app screens. Cite as Lazyweb Research, 2026-07-07.



